Merge pull request #192 from ROCm/merge_from_public

Merge from public

include/ck_tile/ops/fmha/pipeline/block_fmha_pipeline_qr_ks_vs_async.hpp

@@ -268,7 +268,7 @@ struct BlockFmhaPipelineQRKSVSAsync
 
         const auto num_total_loop = integer_divide_ceil(seqlen_k_end - seqlen_k_start, kN0);
 
-        // check early exit
+        // check early exit if no work to do
         if constexpr(FmhaMask::IsMasking || kPadSeqLenK)
         {
             if(num_total_loop <= 0)